The Mediterranean Plastic Soup: synthetic polymers in Mediterranean surface waters

Abstract

The Mediterranean Sea has been recently proposed as one of the most impacted regions of the world with regards to microplastics, however the polymeric composition of these floating particles is still largely unknown. Here we present the results of a large-scale survey of neustonic micro- and meso-plastics floating in Mediterranean waters, providing the first extensive characterization of their chemical identity as well as detailed information on their abundance and geographical distribution. All particles >700 μm collected in our samples were identified through FT-IR analysis (n = 4050 particles), shedding for the first time light on the polymeric diversity of this emerging pollutant. Sixteen different classes of synthetic materials were identified. Low-density polymers such as polyethylene and polypropylene were the most abundant compounds, followed by polyamides, plastic-based paints, polyvinyl chloride, polystyrene and polyvinyl alcohol. Less frequent polymers included polyethylene terephthalate, polyisoprene, poly(vinyl stearate), ethylene-vinyl acetate, polyepoxide, paraffin wax and polycaprolactone, a biodegradable polyester reported for the first time floating in off-shore waters. Geographical differences in sample composition were also observed, demonstrating sub-basin scale heterogeneity in plastics distribution and likely reflecting a complex interplay between pollution sources, sinks and residence times of different polymers at sea.

Figure 1. Size distribution of all particles collected during the survey.

Normalized abundance values (red dots) were obtained by dividing the total number of particles counted in each size class (blue bars) by the width of the respective size bin expressed in millimeters (n = 14,106). Please note that fibers and filaments were not included and that only particles >700 μm were characterized through FT-IR spectroscopy. Secondary vertical axis is in logarithmic scale.

Figure 2. Polymeric composition of all particles >700 μm characterized through ATR FT-IR analysis (n = 4,050 particles).

Values are expressed in percentages. Identification of polymers was performed by comparison with a library of standard spectra and only polymers matching reference spectra for more than 60% were accepted.

Figure 3. Map of the central-western Mediterranean Sea showing the location of all sampling stations and the distribution of un-corrected plastic densities expressed as grams of plastic per km².

Size of the circles is proportional to measured concentration values on a logarithmic scale. Particles <700 μm and synthetic fibers were not included in density calculations. Data were plotted using GPS Visualizer (http://www.gpsvisualizer.com/) and post-edited in Adobe Illustrator CS5. Background map freely retrieved from DEMIS OpenGIS Web Map Server under open copyright licence (http://www.demis.nl/home/pages/wms/docs/OpenGISWMS.htm).

Figure 4. Differences between Adriatic (orange) and western Mediterranean samples (black) in the relative frequencies of the most common types of polymers identified through FT-IR analysis (n = 4,050 particles).

Differences in the mean total abundance (items/m²) and density (g/km²) of particles collected are also shown in the last two panels.

Figure 5. PCA ordination based on the occurrence of the seven most common polymer typologies.

All 74 samples are plotted in relation to the first two principal components (PC1 and PC2). Orange dots represent Adriatic samples (n = 30) while western Mediterranean samples are in black (n = 44). Distance biplot of the eigenvectors (not in scale with data points) is shown to help with the interpretation of the loadings results.

Figure 6. Effect of wind-forcing on the concentration of particles >700 µm.

Wind-induced friction velocities in water () were computed from wind stress (τ) values extracted from ERA-Interim model using density of seawater ρ_w = 1.026 g/cm³. Spearman’s correlation coefficient (r_s) and p value are shown in the top-right corner (n = 74). The dashed line corresponds to cm/s, above which wind correction was applied.